Diodes having a rectifying function are the most fundamental semiconductor elements or components, and various types of diodes are known which have different junction structures.
FIG. 37 is a cross-sectional view showing a pn junction diode 101 having a basic planar-type pn junction. To provide the diode 101, a high-concentration n.sup.+ cathode layer 1 is formed on one of opposite surfaces of a low-concentration n drift layer 2, and a p anode region 3 is formed in a surface layer at the other surface of the n drift layer 2. Cathode electrode 4 and anode electrode 5 are formed in contact with the surfaces of the n.sup.+ cathode layer 1 and p anode region 3, respectively. The diode 101 further includes an oxide film 6 that covers the surface of the pn junction, and a protective film 7 in the form of a nitride film. A p-type peripheral region 8 is formed in a peripheral portion of the pn junction diode 101, and a peripheral electrode 11 is provided on the surface of the peripheral region 8, to extend over a part of the oxide film 6.
The n drift layer 2 is laminated by epitaxial growth on the n.sup.+ cathode layer 1 as a substrate. For example, the impurity concentrations of the n.sup.+ cathode layer 1 and n drift layer 2 are 1.times.10.sup.19 cm.sup.-3, and 1.times.10.sup.15 cm.sup.-3, respectively, and the thicknesses of these layers 1, 2 are 450 .mu.m and 10 .mu.m, respectively. The p anode region 3 is formed by implanting p-type impurities, such as boron ions, using the oxide film 6 as a mask, and thermally diffusing the implanted ions. The p anode region 3 thus formed has a surface impurity concentration of 1.times.10.sup.19 cm.sup.-3, and a diffusion depth of 3 .mu.m.
The graph of FIG. 38 shows a profile of the resistivity measured along a cross section of the pn junction diode 101 of FIG. 37. In FIG. 38, the vertical axis indicates the thickness as measured from the surface of the semiconductor substrate including the n cathode layer 1 and n drift layer 2, and the horizontal axis indicates the resistivity plotted on a logarithmic scale. As shown in the cross section, the diode 101 includes the p anode region 3 having a thickness of 3 .mu.m as measured from the surface of the semiconductor substrate, n drift layer 2 having a thickness of about 60 .mu.m, and the n cathode layer 1 having a low resistivity, which is formed under the n drift layer 2. Generally, the resistivity of a portion of the surface of the p anode region 3 which has the lowest resistance is about 0.01 .OMEGA..multidot.m.
FIG. 39 is a cross-sectional view of a pn junction diode 102 which is a slightly modified example of the planar-type diode of FIG. 37. As in the pn junction diode 101 of FIG. 37, a high-concentration n.sup.+ cathode layer 1 and a low-concentration n drift layer 2 constitute a semiconductor substrate, and a p anode region 3 is formed in a surface layer of the n drift layer 2 of the semiconductor substrate. The pn junction diode 102 is different from the diode 101 of FIG. 37 in that a p ring region 12 having a ring-like shape and a large diffusion depth is formed at the outer periphery of the p anode region 3. While breakdown of the pn junction diode of FIG. 37 is likely to occur in the vicinity of the periphery of the p anode region 3, the p ring region 12 having a larger diffusion depth than the p anode region 3 is formed in the diode of FIG. 39, so as to decrease the gradient of the impurity concentration, thereby to prevent occurrence of the breakdown at around the p anode region 3. As a result, the breakdown occurs uniformly throughout the p anode region 3.
FIG. 40 is a cross-sectional view of a pn junction diode 103 in which p high-concentration regions 13 having a high surface impurity concentration and a large diffusion depth are formed between p anode regions 3 having a low surface impurity concentration and a small diffusion depth, as disclosed in Shimizu et al., IEEE Trans. on Electron Devices ED-31, (1984) p. 1314). When rated current is applied to the diode, the current flows through the p anode regions 3, and therefore the diode exhibits an excellent reverse recovery characteristic. In the reverse bias situation, a depletion layer spreads out from the p high-concentration regions 13, and thus the diode shows a high breakdown voltage. The p high-concentration regions 13 may also serve as the p ring region 12 as described above.
FIG. 41 is a cross-sectional view of a Schottky diode 104 having a basic Schottky j unction. To form the diode 104, a Schottky electrode 15 made of a metal, such as molybdenum, which provides a high Schottky barrier, is formed on a surface of a low-concentration n drift layer 2. A cathode electrode 4 is provided on the rear surface of a n.sup.+ cathode layer 1. A p ring region 12 is formed in a surface layer of the n drift layer 2 so as to surround a contact portion of tile Schottky electrode 15. With the p ring region 12 thus provided, an electric field is prevented from concentrating at the edge of the Schottky electrode 15, and the breakdown voltage of the resulting diode can be increased.
The n drift layer 2 is laminated by epitaxial growth on the high-concentration n.sup.+ cathode layer 1 serving as a substrate. For example, the n.sup.+ cathode layer 1 has a resistivity of 0.004 .OMEGA..multidot.cm, and a thickness of 350 .mu.m, and the n drift layer 2 has a resistivity of 0.90 .OMEGA..multidot.cm, and a thickness of 7 .mu.m.
The graph of FIG. 38 also shows a profile of the resistivity measured along a cross section of the Schottky diode 104 of FIG. 41. The vertical axis indicates the depth as measured from the surface of the semiconductor substrate comprising the n.sup.+ cathode layer and ii drift layer 2, and the vertical axis indicates the resistivity plotted on a logarithmic scale. In the case of a Schottky diode having a breakdown voltage of 60V, for example, the n drift layer 2 having a resistivity of 0.9 .OMEGA..multidot.cm extends from the surface of the semiconductor substrate to a depth of about 7 .mu.m, and the n.sup.+ cathode layer 1 having a resistivity of 0.004 .OMEGA..multidot.cm is formed under the n drift layer 2.
FIG. 42 is a cross-sectional view showing a Schottky diode 105 as a slightly modified example of the Schottky diode 104 of FIG. 41. In the diode 105, trenches 16 are formed in a surface layer of the n drift layer 2, and a Schottky electrode 15 made of molybdenum, for example, is formed on the surface of the n drift layer 2 and the inner walls of the trenches 16. With the trenches 16 thus provided, a contact area of the Schottky electrode 15 is increased, thereby to increase the current capacitance.
FIG. 43 is a cross-sectional view of a composite diode 106 having a pn junction and a Schottky junction. In the composite diode 106, a relatively wide p ring region 12 is formed in a surface layer of an n drift layer 2 so as to surround a contact portion of a Schottky electrode 15, such that the Schottky electrode 15 is in contact with the surface of the p ring region 12 as well as the n drift layer 2, as disclosed in Zettler, R. A. et al.: IEEE Trans. on Electron Devices ED-16, (1969) p. 58. In this case, the p ring region 12 provides a p anode region 3 of a pn junction diode. Thus, the composite diode, in which the pn junction and Schottky junction are combined, provides a low forward voltage when it is forward biased, a high breakdown voltage, and an effect of reducing noise.
FIG. 44 is a cross-sectional view of a composite diode 107 which is a modified example of the composite diode of FIG. 43. In this example, not only the p ring region 12 is formed at the periphery of the n drift layer 2 that contacts with the Schottky electrode 15, but also p anode regions 3 are formed inside the p ring region 12. The Schottky electrode 15 is formed in contact with both exposed portions of the n drift layer 2 and the surfaces of the p anode regions 3, as disclosed in Japanese Patent No. 59-35183. The exposed portions of the n drift layer 2 between the p anode regions 3 have a small width, and a depletion layer spreads out from the p anode regions 3 when a reverse bias is applied, assuring a reduced leakage current.
FIG. 45 is a cross-sectional view of a composite diode 108 which is a slightly modified example of the composite diode 107 of FIG. 44. In this example, a p anode region 3 is formed in a surface layer of the n drift layer 2, and trenches 16 having a larger depth than the p anode region 3 are formed. Further, a Schottky electrode 15 made of molybdenum, for example, is formed in contact with the surface of the p anode region 3 and the inner walls of the trenches 16. In this case, too, the provision of the trenches 15 leads to an increase in the contact area of the Schottky electrode 15, and an increase in the current capacitance.
FIG. 46 is a cross-sectional view of a composite diode 109 which is a slightly modified example of the composite diode 108 of FIG. 45. In this example, trenches 16 are formed in a surface layer of an n drift layer 2, and p anode regions 3 are formed along the inner faces of the trenches 16. Schottky electrode 15 is formed in contact with both a surface layer of the n drift layer 2 where the trenches 16 are not formed, and the surfaces of the p anode regions 3 formed along the inner walls of the trenches (see Kunori, S et al.: Proc. of 1992 Intern. Symp. on Power Semicond. Devices and ICs, Tokyo, (1992) p.69). By providing the trenches 16, and forming the p anode regions 3 on the inner walls of the trenches 16, the leakage current can be reduced in the reverse bias situation.
In the pn junction diodes of FIGS. 37, 39, 40, lifetime killers for accelerating recombination of accumulated carriers are introduced by diffusion of Au or Pt, or irradiation of electron beams, so as to increase the switching speed. However, the introduction of the lifetime killers induces or causes an increase of leakage current. Namely, the leakage current I.sub.R increases if the reverse recovery time t.sub.rr is shortened by introducing a lot of lifetime killers, and the reverse recovery time t.sub.rr, is increased if the leakage current I.sub.R is reduced. Thus, there is a trade-off relationship between the reverse recovery time t.sub.rr and the leakage current I.sub.R. As another problem, the reverse recovery waveform shows hard recovery if a lot of lifetime killers are introduced.
In the Schottky diode of FIGS. 41 and 42, there is a trade-off relationship between the ON-state voltage V.sub.F and the reverse leakage current I.sub.R. The ON-state voltage V.sub.F may be reduced by using a metal having a small barrier height, or lowering the resistance of the n region. In this case, however, the leakage current I.sub.R is undesirably increased in the reverse bias situation. If a metal having a large barrier height is used, or the resistance of the n region is increased, the leakage current I.sub.R is reduced, but the ON-state voltage V.sub.F is increased. Thus, there is a trade-off relationship between the ON-state voltage V.sub.F and the reverse leakage current I.sub.R.
The composite diodes of FIG. 43 through FIG. 46 have aparallel structure of pn junction diode and Schottky diode, and make use of advantages of the respective types of diodes. These composite diodes, however, inherit disadvantages or problems of the pn junction diode and Schottky diodes.
Furthermore, conventional diodes generally have a low ability to withstand avalanche breakdown. In particular, the pn junction has a certain radius of curvature at around a corner portion of the p anode region or p ring region, and therefore the ability to withstand avalanche breakdown is lowered due to concentration of an electric field on the corner portion, as compared with that of the planar pn junction. Thus, the conventional diodes tend to break down due to concentration of current that may result in avalanche breakdown.